What Is OEE? Overall Equipment Effectiveness Explained (2026)

OEE (Overall Equipment Effectiveness) is a manufacturing KPI that measures how effectively a piece of equipment is being used compared to its full potential. It is expressed as a percentage and calculated by multiplying three factors: Availability, Performance, and Quality. A score of 100% OEE means the machine ran whenever it was scheduled, ran at full speed, and produced zero defects.

What Is OEE?

×

Performance Ideal Cycle Time × Count Run Time Speed losses & minor stoppages

×

Quality Good Count Total Count Scrap & rework

=

OEE % ≥ 85% World-class 60–85% Typical

Example: 87.5% × 83.3% × 95.0% = 69.2% OEE — the multiplication effect always reveals more loss than any single factor suggests

Overall Equipment Effectiveness is the single most widely adopted metric in manufacturing for quantifying productive capacity. The concept was formalized by Seiichi Nakajima as part of Total Productive Maintenance (TPM) in the 1960s and remains the standard benchmark for machine and line efficiency in discrete, process, and hybrid manufacturing environments.

OEE answers one clear question: of all the time a machine was scheduled to run, how much of it was truly productive? "Truly productive" means the machine was running, running fast enough, and producing good parts — all three simultaneously. Any shortfall in one factor drags the composite score down.

The formula is:

OEE = Availability × Performance × Quality

Each factor is a ratio between 0 and 1 (or 0–100%). Because they are multiplied together, the composite score is always lower than the weakest individual factor. A machine that is available 90% of the time, running at 90% speed, and producing 90% good parts achieves an OEE of just 72.9% — not 90%.

For the step-by-step math, worked examples, and edge cases around planned versus unplanned downtime, see how to calculate OEE.

Try it: use the free OEE Calculator to get your Availability × Performance × Quality score with world-class benchmarks.

Why OEE Matters in Modern Manufacturing

OEE matters because hidden capacity losses are expensive and frequently invisible without measurement. Most manufacturers know their throughput numbers; far fewer know how much productive time they are leaving on the table inside their scheduled windows.

Key reasons OEE has become a baseline KPI:

• Capital justification — before buying a new machine, OEE reveals whether existing equipment is genuinely constrained or just poorly utilized.
• Continuous improvement targeting — because OEE decomposes into three factors, it tells improvement teams where the loss is occurring, not just that a loss exists.
• Benchmarking — the 85% "world-class" threshold gives teams a universal reference point regardless of industry or machine type.
• Digital transformation foundation — OEE is the most common first use case for IIoT and MES deployments because the data requirements are well-defined and the ROI is measurable.

OEE does not measure everything. It only covers scheduled production time, not the full calendar. It does not capture demand-side underutilization. For those scenarios, TEEP and OOE (covered below) extend the framework.

The Three OEE Factors

Availability

Availability measures the proportion of scheduled time during which the equipment was actually running.

It accounts for all events that stop production when production was expected — unplanned breakdowns, planned maintenance performed during scheduled production time, changeovers, material shortages, and operator-related stoppages.

Availability = Run Time ÷ Planned Production Time

A machine scheduled for 8 hours that suffers 1 hour of downtime has an Availability of 87.5%.

Availability is the factor most directly influenced by maintenance strategy. Predictive maintenance programs driven by PLC sensor data target this factor specifically by eliminating unexpected breakdowns before they occur.

Performance

Performance measures how fast the equipment ran compared to its theoretical maximum speed during the time it was actually running.

It captures speed losses: equipment running slower than its ideal cycle time, and brief micro-stops (typically under 5 minutes) that do not trigger a formal downtime record but still erode output.

Performance = (Ideal Cycle Time × Total Count) ÷ Run Time

A machine with a 30-second ideal cycle time that produces 800 parts in a 480-minute run has a Performance of 83.3%.

Performance losses are often the least visible category. Operators may throttle machine speed to reduce rejects or avoid a recurring fault — each individual instance is small, but the cumulative effect over a shift is significant.

Quality

Quality measures the proportion of total output that meets specification on the first pass, without rework.

It distinguishes between parts that were good on the first attempt and those that required rework or were scrapped entirely.

Quality = Good Count ÷ Total Count

A machine that produces 800 parts but scraps 40 of them has a Quality rate of 95%.

Quality losses are directly tied to process capability, tooling condition, raw material variation, and operator settings. In automated lines, PLC reject counters and vision system outputs feed this figure directly.

Putting the Three Factors Together

Factor	Example Value	Contribution
Availability	87.5%	Scheduled 8 h, ran 7 h
Performance	83.3%	Ran at 83% of ideal speed
Quality	95.0%	95% first-pass yield
OEE	69.2%	0.875 × 0.833 × 0.95

The multiplication effect means that even three individually respectable numbers can combine into a composite score that reveals substantial room for improvement.

The Six Big Losses

The Six Big Losses framework, also from Nakajima's TPM methodology, categorizes every form of OEE loss into one of three pairs — one pair per factor. Understanding which loss category is dominant tells the improvement team where to focus.

Availability Losses

1. Unplanned Downtime (Breakdowns) Equipment failures that were not anticipated: motor trips, sensor faults, pneumatic failures, tooling breakage. These are the highest-visibility losses and the primary target of preventive and predictive maintenance programs.

2. Planned Downtime (Changeovers and Setup) Time deliberately taken for product changeovers, cleaning, planned maintenance, and tooling changes. SMED (Single Minute Exchange of Die) methodology targets this loss. Note: some OEE implementations exclude planned maintenance from the scheduled production window entirely, which changes the Availability calculation — define your boundaries clearly before comparing sites.

Performance Losses

3. Small Stops (Minor Stoppages) Interruptions under a defined threshold (commonly 5 minutes) that do not get logged as formal downtime events: jams, sensor false-trips, manual interventions to clear a part, brief conveyor blockages. These losses are systematically underreported in manual tracking systems and are only reliably captured by automated, PLC-sourced OEE collection.

4. Reduced Speed (Slow Cycles) The machine runs continuously but below its ideal cycle time. This includes intentional speed reductions (running "safe" to prevent quality defects) and gradual degradation from worn tooling, bearing friction, or pneumatic pressure loss.

Quality Losses

5. Production Rejects (Defects) Parts produced during normal, stable production that fail to meet specification. Scrap and rework both count against Quality, though rework parts are sometimes tracked separately to distinguish recoverable losses from unrecoverable ones.

6. Startup Rejects (Yield Losses) Parts produced during the warm-up or startup period before the process stabilizes: initial rejects after a changeover while settings are dialed in, first-off parts from a cold start, and rejects produced immediately after a fault clearance while the process re-stabilizes.

Six Big Losses Summary Table

Loss	OEE Factor	Typical Root Causes
Unplanned downtime	Availability	Equipment failures, electrical faults
Planned downtime	Availability	Changeovers, scheduled maintenance
Small stops	Performance	Jams, sensor trips, manual interventions
Reduced speed	Performance	Worn tooling, intentional throttling
Production rejects	Quality	Process variation, material defects
Startup rejects	Quality	Warm-up, post-changeover instability

Six Big Losses — Mapped to OEE Factors

AVAILABILITY Run Time ÷ Planned Time Loss 1: Unplanned Downtime Motor trips, sensor faults tooling breakage Loss 2: Planned Downtime Changeovers, SMED targets scheduled maintenance PERFORMANCE Ideal Cycle × Count ÷ Run Time Loss 3: Small Stops Jams <5 min, sensor false-trips invisible to manual logs Loss 4: Reduced Speed Running below ideal cycle time worn tooling, intentional throttle QUALITY Good Count ÷ Total Count Loss 5: Production Rejects Scrap & rework during stable production Loss 6: Startup Rejects Post-changeover warm-up rejects before process stabilises

What Is a Good OEE Score?

OEE scores are interpreted against three widely accepted reference benchmarks:

OEE Score	Interpretation
≥ 85%	World-class. The benchmark established by Nakajima for high-performing discrete manufacturing operations. Sustained 85%+ requires mature TPM and tight process control.
60–85%	Typical for manufacturing operations that have some improvement programs in place. Common in mid-size discrete manufacturers. Significant improvement potential exists.
< 40–60%	Common starting point for operations with little formal OEE tracking. Often a surprise to management — the losses were there before, just unmeasured.

Important context for interpreting these benchmarks:

• The 85% world-class figure originates from Nakajima's discrete manufacturing context. Process industries (chemicals, food, pharmaceuticals) often operate at lower OEE by design because changeover-heavy or cleaning-intensive processes structurally limit Availability and Performance.
• A single OEE number hides composition. Two plants at 65% OEE may have entirely different loss profiles — one dominated by breakdowns, the other by speed losses — and require completely different improvement interventions.
• OEE should be trended over time and benchmarked internally before external comparisons are made. A consistent 68% with an upward trajectory is more valuable than a volatile 72%.
• New to OEE measurement? Expect your first measured score to be lower than intuition suggests. Manual tracking typically overcounts good production and undercounts stop events. Automated, PLC-sourced OEE consistently reveals losses that paper-based systems miss.

OEE vs TEEP vs OOE

OEE is one of three related equipment effectiveness metrics. They differ in the denominator — what time base is used.

Metric	Full Name	Time Base	What It Measures
OEE	Overall Equipment Effectiveness	Planned production time	Effectiveness within scheduled windows
TEEP	Total Effective Equipment Performance	All calendar time (24/7/365)	Utilization of total theoretical capacity
OOE	Overall Operations Effectiveness	Operating time (incl. unscheduled)	Effectiveness including unscheduled runs

When to use each:

• OEE is the right metric for measuring continuous improvement on the shop floor. It focuses on the time that was planned for production, making it actionable for operators and maintenance teams.
• TEEP is the right metric for strategic capacity planning. An OEE of 85% during a single-shift operation translates to a TEEP of roughly 35% when measured against 24/7 calendar time — revealing headroom for additional shifts.
• OOE sits between the two and is used when operations regularly run unscheduled overtime or weekend shifts that should be captured in the effectiveness calculation.

For most shop-floor improvement programs, OEE is the primary metric. TEEP becomes relevant when evaluating whether to add capacity or add shifts.

How OEE Is Measured and Collected

Manual OEE Tracking

The simplest OEE implementations use paper logs or spreadsheets. Operators record start and stop times, reason codes for stoppages, and part counts at the end of each shift. A supervisor or engineer compiles the data.

Limitations of manual tracking:

• Small stops (< 5 minutes) are almost never recorded — operators do not interrupt the line to log a 90-second jam
• Reason codes are assigned retrospectively and are often inaccurate
• Data entry errors and missing records are common
• Results are available hours or days after the events occurred, limiting real-time response
• Improvement teams cannot trust the data enough to make confident root-cause decisions

Manual tracking is a reasonable starting point for organizations with no OEE baseline, but it will systematically underreport Performance losses and overstate Availability.

Automated OEE from PLC and Automation Layer Data

The most accurate and actionable OEE systems capture data directly from the PLC and automation layer, eliminating manual data entry entirely. This approach is the standard in Industry 4.0 implementations and is the method that makes continuous, real-time OEE monitoring practical at scale.

How PLC data maps to OEE inputs:

OEE Input	PLC/Automation Data Source
Machine running / stopped	Machine state bit (Running, Faulted, Idle, Changeover)
Fault reason	Fault code register or alarm tag
Part count (total)	High-speed production counter (encoder or photoeye pulse)
Good part count	Counter decremented by reject signal from vision system or quality gate
Reject count	Reject output tag from vision system, checkweigher, or manual reject station
Ideal cycle time	Configured constant in HMI or SCADA recipe management
Changeover start/end	Mode selector or recipe change event

Modern PLCs log machine-state transitions with millisecond-resolution timestamps. This means that even a 45-second jam — invisible to a paper log — is captured as a discrete stop event with an accurate duration and the fault code that caused it.

The data collection architecture typically looks like this:

1. PLC layer — machine state bits, counters, and fault codes are maintained in the controller as part of normal control program execution. No additional PLC programming is required beyond ensuring the right tags exist and are updated correctly.
2. Edge layer / SCADA — an OPC UA or MQTT-enabled edge gateway (or the SCADA system itself) polls or subscribes to the relevant PLC tags at a defined scan rate (commonly 1–5 seconds for state data, faster for counters on high-speed lines). See IIoT and PLC integration for communication protocol options.
3. MES or OEE software — the collected tag data is aggregated into shift-level OEE calculations. The MES system applies the business rules: what constitutes planned versus unplanned downtime, what the ideal cycle time is for the current product, and how rejects are classified. See MES vs ERP for how this layer fits into the broader enterprise architecture.
4. Dashboards and reporting — real-time OEE dashboards update continuously, enabling operators and supervisors to see loss events as they happen rather than reviewing them the following morning.

ISA-95 machine states provide a standardized vocabulary for state-based OEE collection. The core states relevant to OEE are: Producing (mapped to Run Time), Standby (scheduled but not producing — feeds Availability loss), Maintenance (planned downtime), and Faulted (unplanned downtime). Mapping your PLC state bits to ISA-95 states before implementing OEE ensures your data is compatible with standard MES and OEE platforms.

Manual vs Automated OEE: A Direct Comparison

Dimension	Manual Tracking	Automated (PLC-sourced)
Small stop capture	Rarely captured	Every event recorded
Data latency	Hours to days	Real-time
Reason code accuracy	Retrospective / subjective	Fault code from controller
Operator burden	High	None
Data trust for CI decisions	Low	High
Implementation cost	Low	Medium to high
Scalability across lines	Poor	Excellent

For operations running more than a handful of machines, automated OEE collection pays for itself rapidly through the improvement actions it enables. The manufacturing automation guide covers the broader automation infrastructure that supports this kind of data capture.

OEE and Predictive Maintenance

OEE Benchmark Bands — What Your Score Means

0% 20% 40% 60% 80% 100%

New to OEE Below 40% — First measured scores; losses were always there, now visible Starting point 40–60% — Common in operations new to formal OEE measurement Typical mfg. 60–85% — Mid-size manufacturers with active improvement programs World- class ≥ 85% — Nakajima world-class benchmark (mature TPM + tight process control) 85%

Context: two plants at 65% OEE may need completely different interventions — always review the three individual factors alongside the composite score

OEE and predictive maintenance are complementary disciplines. OEE identifies that a machine is losing Availability or Performance; predictive maintenance determines why and anticipates when the next failure will occur.

A degrading bearing, for example, will first appear in OEE data as a gradual Performance loss (the machine slows slightly as friction increases) before it appears as an Availability loss (the bearing fails completely and the machine stops). Predictive maintenance systems monitoring vibration, temperature, and current draw from the same PLC data infrastructure can catch the degradation trend and schedule a planned replacement — converting an unplanned Availability loss into a controlled Planned Downtime event.

PLC-based predictive maintenance covers the sensor types, data collection methods, and analytics approaches used to close this loop between OEE visibility and proactive maintenance action.

---

Frequently Asked Questions

What is OEE?

OEE stands for Overall Equipment Effectiveness. It is a manufacturing KPI that measures how productively a machine or production line is being used relative to its full potential during scheduled production time. OEE is calculated as: Availability × Performance × Quality. A score of 100% means the machine ran all scheduled time, at full speed, with zero defects. In practice, world-class OEE is considered to be 85% or above.

What does OEE stand for?

OEE stands for Overall Equipment Effectiveness. The term was coined by Seiichi Nakajima as part of the Total Productive Maintenance (TPM) methodology developed in Japan in the 1960s. It became an internationally recognized standard metric through the growth of lean manufacturing and TPM programs globally.

What is a good OEE score?

The widely accepted benchmarks are:

• 85% and above — world-class, the target for high-performing discrete manufacturing operations
• 60–85% — typical for manufacturers with active improvement programs
• 40–60% — common starting point for operations new to OEE measurement

Context matters. A 75% OEE with an upward trend and well-understood loss categories is more valuable than a one-time 85% reading with no supporting analysis. Always review the three individual factors alongside the composite score.

What are the Six Big Losses?

The Six Big Losses are the six categories of OEE loss defined in TPM methodology:

1. Unplanned downtime (Availability) — unexpected equipment failures and breakdowns
2. Planned downtime (Availability) — changeovers, setups, and scheduled maintenance performed during production time
3. Small stops (Performance) — brief interruptions under ~5 minutes that do not get formally logged
4. Reduced speed (Performance) — machine running slower than its ideal cycle time
5. Production rejects (Quality) — defective parts produced during stable running
6. Startup rejects (Quality) — defective parts produced during warmup or re-stabilization after a changeover or fault